card: correct the calibration note (AUROC definitions), add the frozen-backbone comparison on long text

README.md

@@ -47,10 +47,11 @@ general item, 150 to 190 ms per screenshot with all of a step's questions, on on
 
 ## Known limits
 
-- Out of its training domains it is over-confident: on JevBench's public hard tier (text-only long documents) it is
-  52% accurate at 91% mean confidence. Yes/no ranking on general images (AUROC 0.84 to 0.91) is below the frozen
-  backbone's (0.96). A temperature fitted on out-of-domain data is the intended fix; V10 tests a readout that keeps
-  the backbone's yes/no geometry.
+- On text-only long documents (JevBench's public hard tier) it is 52% accurate at 91% mean confidence: badly
+  over-confident out of domain. The frozen backbone read from next-token logits is 59% at ECE 0.35 on the same items,
+  so most of this is the backbone's own behaviour on long text; a temperature fitted on out-of-domain images
+  (T = 1.3 on GQA val) moves the hard-tier ECE from 0.28 to 0.24 only. Do not use the stated probabilities on long
+  text documents as calibrated confidences. On general images the yes/no class AUROC is 0.97 to 0.98.
 - Web grounding on Mind2Web-style pages 0.823 and macOS desktop grounding 0.822 are below the earlier grounding-only
   checkpoint (0.850 / 0.914).
 - Trained and evaluated on 1280x1000 desktop web screenshots; not on mobile, logged-in or form-heavy flows.